FMEA: Failure Mode and Effects Analysis in Practice
Every quality failure has a history. Somewhere in the process — in the design, the materials, the assembly steps, or the handling instructions — there was a warning sign that nobody caught in time. Failure Mode and Effects Analysis (FMEA) is the discipline that reads those warning signs before the failure occurs, not after the warranty claim arrives.
What Is FMEA?
FMEA is a structured, team-based technique for identifying every way a product or process can fail, analysing the consequence of each failure, and prioritising corrective actions before those failures reach the customer. It is a core Lean Six Sigma tool appearing in automotive (AIAG/VDA), aerospace (AS9100), healthcare, and construction quality standards. The output is a living document — the FMEA table — that captures each failure mode alongside its risk score and drives teams to invest improvement energy where risk is highest.
The Three Dimensions: Severity, Occurrence, Detection
Every row in the FMEA table is scored on three dimensions, each rated from 1 to 10:
Severity (S) — how bad is the consequence if this failure mode reaches the customer? A score of 1 means the customer would not notice; a score of 10 means a safety hazard or regulatory non-compliance.
Occurrence (O) — how likely is the root cause of this failure mode to materialise? A score of 1 means the cause is extremely unlikely; a score of 10 means it occurs routinely.
Detection (D) — how likely are existing controls to catch the failure before it reaches the customer? A score of 1 means detection is almost certain; a score of 10 means the failure is virtually undetectable.
The Risk Priority Number (RPN)
The Risk Priority Number is calculated as RPN = S × O × D. It produces a single number on a scale of 1 to 1,000 that combines all three risk dimensions into one priority score. A common threshold for mandatory action is RPN ≥ 100 or ≥ 120, though thresholds vary by industry and organisational risk appetite.
One important caution: do not rely on the RPN alone. Any failure mode where Severity = 9 or 10 demands immediate attention regardless of the RPN — because a catastrophic consequence is never acceptable, even if it seems unlikely or appears detectable. High-severity items should be addressed first, before the RPN-ranked list.
Building the FMEA Table
A practical FMEA table includes the following columns for each item in scope:
Item / Function — what the component or process step is supposed to do.
Failure Mode — the way it could fail to perform that function.
Effect of Failure — the consequence to the customer or system.
Severity (S) — rated 1–10; 10 = safety hazard or regulatory non-compliance.
Cause of Failure — the root mechanism. Be specific: vague causes produce vague actions.
Occurrence (O) — rated 1–10; likelihood the cause will materialise.
Current Controls — prevention or detection measures already in place.
Detection (D) — rated 1–10; 1 = almost certain to catch; 10 = virtually undetectable.
RPN = S × O × D.
Recommended Action / Owner / Target Date — for high-RPN and high-severity items.
Revised RPN — re-scored after the action is implemented.
Process FMEA vs. Design FMEA
There are two primary types of FMEA in industrial practice. Process FMEA (PFMEA) analyses a manufacturing or service delivery process — each step, each machine, each handoff. Design FMEA (DFMEA) analyses a product design before it is built, identifying how geometry, materials, tolerances, or interfaces could fail. The two are complementary: a DFMEA may reveal a feature that is inherently difficult to manufacture consistently, and the PFMEA then addresses the process controls needed to manage it. Complex products benefit from both.
Prioritising Actions: High Severity First
Prioritise in this sequence: first, address all Severity = 9 or 10 items regardless of RPN; second, work through high-RPN items in descending order. Actions that reduce Severity are most valuable — they eliminate the harm at the source. Actions that improve Detection are least preferred, because they accept that the failure will still occur and merely try to catch it first.
Updating the FMEA After Improvements
An FMEA is not a one-time exercise. After implementing corrective actions, the team re-scores the affected rows. If the revised RPN falls below threshold and no high-severity items remain unmitigated, the action is considered effective. The FMEA table is then archived as a living quality record — consulted whenever the product or process changes or a new failure mode is discovered in the field.
Applications Across Industries
FMEA is not limited to automotive assembly lines. In healthcare, process FMEA analyses medication administration workflows where the stakes are patient safety. In construction, it is applied to critical installation sequences where a single missed step can have life-safety consequences. In IT, FMEA identifies single points of failure in system architectures and release processes before they cause production outages. Wherever it is applied, FMEA converts tacit knowledge — the expertise experienced team members carry in their heads — into an explicit, auditable risk register that survives staff turnover.
XNM Consulting helps organisations embed structured quality tools like FMEA into their operations and project delivery. Learn more on our Strategic Advisory page.